Novelty Production and Evolvability in Digital Genomic Agents: Logical Foundations and Policy Design Implications of Complex Adaptive Systems

Abstract

Novelty, innovation and surprises are ubiquitous in evolution and human social systems, with the adaptive immune system (AIS) and capitalism being respective exemplars. While novelty production embodies the hallmark of complex adaptive systems (CAS), as yet there is no consensus on what produces the ‘smarts’ that allows for innovative behaviours. Markose (2017) has shown how the Godel-Turing-Post (G-T-P) digital machinery necessary to encode the Godel sentence, is instrumental in genomic novelty and human proteanism. With evidence that genomic evolution acquired the mirror systems with the AIS and also in cognitive mirror neurons, the necessary G-T-P conditions for Self-Ref (self-reference) and Self-Rep (self-representation) were in place to encode the Liar qua hacker in the Godel sentence. The latter allows a digital entity to self-report it is under attack and it becomes the trigger for the elimination of the Liar/hacker. Failing that, the Nash equilibrium entails a co-evolutionary arms race of innovations from which unilateral withdrawal spells failure. This represents undecidable structure changing dynamics wherein predictable formal rules may suffer demise. The idea that predictability can lead to system failure and policy ineffectiveness has a long provenance in Kantian principles based end-neutral coercive laws, and in the more recent intuitively held positions of the Lucas Critique on the need for surprises and also Goodhart's Law. The serial collapse of currency pegs and the massive system failure in the 2007 Great Financial Crisis were the consequence of a paradigmatic blind spot: pre-commitment to formal rules by authorities effectively eliminated 'surprises' on their own part by withdrawing from the co-evolutionary arms race and gave rule breakers (the Liar) free reign to game the system.